> Markdown version of [/jobs/ext/4754314-software-engineer](https://www.wearedevelopers.com/jobs/ext/4754314-software-engineer). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md). --- # Software Engineer - **Company:** EVB - **Location:** United Kingdom (Remote available) - **Salary:** £104,000 - £208,000 - **Contract:** Contract - **Skills:** Java (Programming Language), Artificial Intelligence, C++ (Programming Language), Data Structures, Software Debugging, Python (Programming Language), Machine Learning, Software Engineering, TypeScript, Reinforcement Learning, Rust (Programming Language), Golang - **Published:** November 17, 2026 - **Apply:** https://www.apply4u.co.uk/jobs/software-engineer/56995279 ## About the Role Significant hands-on expertise in at least one of: Python3, Java, Rust, Go, C++, or TypeScript. Deep understanding of algorithms, data structures, and software engineering principles. Proven ability to debug complex systems, resolve bugs, and deliver effective optimizations. Experience with large codebase refactoring and legacy system modernization. Track record of delivering high-impact features from conception through delivery. Strong documentation and communication skills to articulate technical concepts clearly. Interest in AI and its technical challenges - previous AI experience not required. ## Description You'll create Reinforcement Learning Environments that test a model's ability to solve real software engineering problems: tracking down subtle bugs, building non-trivial features, refactoring, and optimizing performance. Each task means defining the problem, building a reproducible environment, and writing a reference solution. Scope of Work Create realistic, challenging coding tasks for frontier AI models, drawn from repositories you know well. Design problems that require real engineering skill - subtle bugs, non-trivial features, meaningful performance work - not lookup or boilerplate. Build a reproducible environment for each task. Write deterministic verifiers that accept correct solutions across different valid implementations, while rejecting incorrect, incomplete, or hard-coded approaches. Provide a reference solution showing how the task can be solved.